PeopleofSuperintelligence

Researcher

Naman Goyal

Naman Goyal is a researcher in large language model pretraining whose coauthored papers include RoBERTa, mBART, OPT and LLaMA. He studied computer science at Georgia Tech, worked as a software engineer at Yelp, and then spent 2018 to 2025 as a researcher at Meta. During those years he coauthored RoBERTa in 2019, mBART in 2020, OPT in 2022 and LLaMA in 2023, and he was first author of the 2021 paper that scaled XLM-R to 10.7B parameters. Since 2025, he is a member of technical staff at Thinking Machines Lab.

Badges

Timeline

  1. 2008–2016

    Pune and a Georgia Tech master's

    He studies computer engineering at Savitribai Phule Pune University from 2008 to 2012, then pursues a master's in computer science at Georgia Tech from 2014 to 2016. There he serves as a graduate teaching assistant, grading and writing autograder scripts for a graduate-level artificial intelligence class.

  2. 2016–2018

    Yelp and a summarization paper

    He works as a software engineer at Yelp from 2016 to 2018. In 2016, he and Jacob Eisenstein publish A Joint Model of Rhetorical Discourse Structure and Summarization in the proceedings of the Workshop on Structured Prediction for NLP, held in Austin.

  3. 2018–2025

    Researcher at Meta

    From 2018 to 2025, he works at Meta as a researcher. At Facebook AI Research, his focus is large language models, natural language processing and computer vision.

  4. 2019–2020

    RoBERTa and mBART

    In 2019, he is a coauthor of RoBERTa, a replication study of BERT pretraining whose best model reaches state-of-the-art results on GLUE, RACE and SQuAD. In 2020, he coauthors mBART, a sequence-to-sequence denoising auto-encoder pre-trained on monolingual corpora in many languages.

  5. 2021

    Scaling XLM-R to 10.7B parameters

    In 2021, he is first author of Larger-Scale Transformers for Multilingual Masked Language Modeling, published in the proceedings of the Workshop on Representation Learning for NLP. The paper presents multilingual models with 3.5B and 10.7B parameters, XLM-R XL and XLM-R XXL, which outperform XLM-R by 1.8% and 2.4% average accuracy on XNLI.

  6. 2022

    The OPT open models

    In 2022, he is the third author of OPT: Open Pre-trained Transformer Language Models, a suite of decoder-only transformers from 125M to 175B parameters. The team reports that OPT-175B is comparable to GPT-3 while requiring 1/7th the carbon footprint to develop.

  7. 2023

    LLaMA foundation models

    In 2023, he is one of the 14 authors of LLaMA, a collection of foundation language models from 7B to 65B parameters trained on publicly available datasets exclusively.

  8. Since 2025

    Technical staff at Thinking Machines

    In 2025, he joins Thinking Machines Lab as a Member of Technical Staff, where he continues multimodal research.

Career

  1. Since 2025Thinking Machines LabMember of Technical Staff
  2. 2018–2025FAIR (Meta)Researcher
  3. 2016–2018YelpSoftware Engineer
  4. 2014–2016Georgia Institute of TechnologyGraduate Teaching Assistant

Education

  1. 2016Georgia Institute of TechnologyMaster's in computer science
  2. 2012Savitribai Phule Pune University

From the biography.

  1. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Naman Goyal. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.
  2. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Naman Goyal. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  3. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Thinking Machines Lab. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  4. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes Thinking Machines Lab. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Naman Goyal yet. Send us a free-licensed photo.

  1. Naman Goyal (Thinking Machines Lab) — Lacuna lacuna.tiptreesystems.com. Checked 21 October 2026.
  2. Naman Goyal | alphaXiv alphaxiv.org. Checked 21 October 2026.
  3. A Joint Model of Rhetorical Discourse Structure and Summarization - ACL Anthology aclanthology.org. Checked 21 October 2026.
  4. [1907.11692] RoBERTa: A Robustly Optimized BERT Pretraining Approach arxiv.org. Checked 21 October 2026.
  5. [2001.08210] Multilingual Denoising Pre-training for Neural Machine Translation arxiv.org. Checked 21 October 2026.
  6. Larger-Scale Transformers for Multilingual Masked Language Modeling - ACL Anthology aclanthology.org. Checked 21 October 2026.
  7. [2205.01068] OPT: Open Pre-trained Transformer Language Models arxiv.org. Checked 21 October 2026.
  8. [2302.13971] LLaMA: Open and Efficient Foundation Language Models arxiv.org. Checked 21 October 2026.
  9. Meta LLAMA team exodus:Meta Loses Top LLAMA Researchers - Kingy AI kingy.ai. Checked 21 October 2026.